Impact of satellite constellations on astronomical observations with ESO telescopes in the visible and infrared domains
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Astronomy & Astrophysics manuscript no. output c ESO 2020 March 3, 2020 Impact of satellite constellations on astronomical observations with ESO telescopes in the visible and infrared domains Olivier R. Hainaut1 and Andrew P. Williams1 European Southern Observatory, Karl-Schwarzschild-Strasse 2, D-85748 Garching bei München, Germany Submitted 2020-Jan-14 AA/2020/37501 – Accepted 2020-Feb-28 ABSTRACT The effect of satellite mega-constellations on astronomical observations in the visible, near-infrared, and thermal infrared domains is estimated using a simple methodology, which is applied to ESO telescopes and instruments as examples (radio and (sub-)millimetre domains are not considered here). The study considers a total of 18 constellations in development by SpaceX, Amazon, OneWeb, and others, with over 26 thousand satellites, constituting a representative distribution. This study uses a series of simplifications and assumptions in order to obtain conservative, order-of-magnitude estimates of the effects: the satellites are assumed to be uniformly spread over the Earth’s globe, and their magnitude is estimated using a simplistic model calibrated on actual observations. The effect on various types of ground-based telescopic observations is estimated using a geometric probabilistic approach. The ‘trains’ of very-low altitude satellites typically observed immediately after launch are extremely bright due to their very low orbit. They also fall very quickly in the shadow of the Earth after sunset. However, this initial bright state is not considered further, as the satellites quickly disperse into their higher altitude orbits. The number of illuminated satellites from the constellations above the horizon of an observatory ranges from approximately 1600 immediately after sunset, decreasing to 1100 at the end of astronomical twilight, most of them (∼ 85%) close to the horizon (below 30◦ of elevation). The large majority of these satellites will be too faint to be seen with the naked eye: at astronomical twilight, 260 would be brighter than magnitude 6 (i.e. visible in exceptional conditions), 110 brighter than 5 (i.e. visible in good conditions). Again, most of them (∼ 95%) will be close to the horizon (below 30◦ of elevation). The number of naked-eye satellites plummets as the Sun reaches 30–40◦ of elevation below the horizon. Specular flares and occultations by satellites are expected to cause only negligible effects on telescopic astronomical observations. The light trail caused by the satellite would ruin a small fraction (below the 1% level) of telescopic exposures using narrow to normal field imaging or spectroscopic techniques in the visible and near-infrared during the first and last hours of the night. Similarly, the thermal emission of the satellite would affect only a negligible fraction of the observations in the thermal infrared domain. However, wide-field exposures and long medium-field exposures would be affected at the 3% level during the first and last hours of the night. Furthermore, ultra-wide imaging exposures on a very large telescope (where saturation of the satellite trails has a ruinous effect on the detectors, such as those from the National Science Foundation’s Vera C. Rubin Observatory, formerly known as LSST), would be significantly affected, with 30 to 40% of such exposures being compromised during the first and last hours of the night. Coordination and collaboration between the astronomical community, satellite companies, and government agencies are therefore critical to minimise and mitigate the effect on astronomical observations, in particular on survey telescopes. Key words. Astronomical instrumentation, methods and techniques; Light pollution; Methods: observational; Site testing 1. Introduction tions. It does not replace careful, detailed simulations taking into account the intricacies of the orbital distribution, the complexity While artificial satellites have until now been a substantial of estimating the brightness of a satellite, and so on, but provides concern for radio astronomy and a relatively minor issue for a first, quantitative estimate of the effect. As the simplifications observers in the optical spectrum, the recent launches of the and assumptions used are conservative, the estimated effect is SpaceX Starlink constellation with their spectacularly bright likely greater than the actual effect. post-launch appearance, and the growing publicity of the plans A series of known constellations were taken into account, of other companies for major constellations of thousands of totalling 18 (sub-) constellations and over 26 000 satellites. A satellites, have caused alarm in the astronomy community. The simple approximation was used to model their distribution. issue has also generated substantial media attention, which has The effect on visible and near-infrared observations (NIR) highlighted the emotional and moral dimensions of the issue was estimated using a simple model for their brightness, which that go beyond the impacts on astronomical science. In order for computes their magnitude as a function of their orbital altitude the astronomy community to respond to these developments and and of the angular elevation above the horizon. While this model work constructively with industry, funding agencies, and regu- is extremely crude, it is calibrated using known satellites and lators, a factual and quantitative assessment of the impacts is observations of SpaceX’s recently launched ‘Starlink’ satellites, required. and is validated by direct photometric observations of a Star- This paper aims at quantifying the effect of large satellite link satellite. Additionally, an estimate of the number of specu- constellations on visible, near-infrared (NIR) and thermal IR as- lar reflections – the bright satellite flares – is provided by scal- tronomical observations using a series of simplifying assump- ing the flares observed for the Iridium first-generation satellites. Article number, page 1 of 12
A&A proofs: manuscript no. output 1984), that is, all satellites from a constellation are on similar cir- Density of Sat per angular area 200 cular orbits with the same altitude and the same high inclination, i=50.0deg grouped on a series of orbital planes whose nodes are uniformly 180 i=80.0deg Uniform distributed along the equator. The inclinations of the constel- 160 lations considered are in the range of 42 to 80◦ . However, in Density [arbitrary unit] 140 what follows, the actual distribution is simplified: the satellites are assumed to be uniformly distributed over the whole Earth. 120 The actual Walker constellation distribution causes a dearth of 100 satellites in the polar regions and an increase of the number of satellites at latitudes close to the inclination of the constellation. 80 This approximation will therefore overestimate the number of 60 satellites above the equatorial and low-latitude regions, will un- 75 50 25 0 25 50 75 derestimate this number at latitudes close to the orbital inclina- Latitude [deg] tion, and will overestimate again the number in regions above very high-latitudes, as illustrated in Fig. 1. Many of the large pro- Fig. 1. Density of satellites on their orbital sphere as a function of lat- fessional telescopes are at latitudes lower than ∼ 30◦ ; for them, itude, for two Walker constellations with orbital inclinations of 50 and this approximation is conservatively overestimating the number 80◦ and for the uniform approximation used in this paper. The total of satellites. In this simplified configuration, the latitude of an number of satellites is the same in all cases. observatory has no effect on the number of satellites that affect it. Accounting for all of the above, the effect is computed on vari- ous types of observations: different exposure duration, various 2.3. Number of satellites in range field-of-view sizes, visible and thermal-infrared observations, and also occultation by a non-illuminated satellite passing in We now estimate the number of satellites above the horizon of an front of the object observed. Computation was performed us- observatory (zenithal distance z = 90◦ ), and above z = 60◦ (ele- ing a geometric probabilistic approach –what fraction of the sky vation e = 60◦ above the horizon, also corresponding to an air- would be contaminated by satellites for the considered obser- mass of 2, the limit above which most of the astronomical obser- vation type– rather than a direct simulation involving repeated vations are performed). At this stage, we consider only whether modelled observations. the satellite is in range, independently of it being illuminated or The effects on millimetre and submillimetre observations are not. not considered here. A separate paper will estimate these effects. To do this, we compute the area of the spherical cap above This paper focuses on the effect on pointed observations; other the threshold value of z; the number of satellites present in that science cases, in particular wide-field observations such as sur- cap is then approximated by the ratio of the cap area to the area of veys, could be more affected. Also not considered are the im- the sphere, multiplied by the number of satellites on the sphere. pacts on the amateur astronomy and astrophotography commu- Here, we introduce γ, the orbital position angle, measured nity. The effects on observatory operations, associated cost im- between the satellite and the zenith measured at the centre of the plications, and political or regulatory issues are beyond the scope Earth. The condition of visibility of a satellite above the horizon, of this paper. z < 90◦ , converts into γ < γo , with ! REarth 2. Constellation and number of satellites γo = arccos(REarth /RSat ) = arccos , (1) REarth + h 2.1. Known upcoming constellations with orbits where rEarth = 6375 km is the radius of the Earth, rSat = rEarth + Table 1 lists publicly known future mega-constellations that are h is the radius of the satellite’s orbit, and h the altitude of the in development, the number of satellites that are planned for satellite above the Earth. The value of γo is reported in Table 1 launch, and the orbital altitude h of the satellites. The list is for the various constellations. To generalise Eq. 1 to any value of neither complete nor accurate, as it is based on generic web z, we first introduce the angles β = π/2 − z (angle OŜH, opposite searches, operator websites, and on official documents submit- to z in the right triangle OHS; see Fig. 2) and δ (angle OŜC). ted to the Federal Communications Commission (FCC). Some From the sinus theorem applied to triangle OCS, of the constellations have already been cancelled. Other, non- telecommunication constellations have filed documentation with sin δ sin (π − z) = , (2) the FCC. Overall, this list should therefore be considered only REarth RSat as a representative list of a variety of constellations and a large or number of satellites, rather than an exact representation of what will be launched in the coming years, which is sufficient for this REarth sin z sin δ = . (3) study. Its results can be scaled to smaller or larger constellations. RSat The original Iridium constellation is also included, although its size and effects are negligible compared to the others. It is used Summing the angles of the triangle OSC, for calibration of the flare numbers. π = γ + δ + (π − z) . (4) 2.2. Distribution of the satellites Using Equations 1 and 4, ! The satellites in these mega-constellations will likely be organ- REarth ised in configurations similar to the Walker constellation (Walker γ = z − arcsin sin z . (5) RSat Article number, page 2 of 12
Hainaut O. R.: Satellite Constellations Constellation a Satellites Satellites Alt. for above horizon above zmax = 60◦ Orbital Magnitude at Zenith Period Zenith z = 60◦ N h γo a % N a % N P [km] [deg] [deg] [deg] [h] [mag] [mag] 1 2 3 4 5 6 7 8 9 10 11 12 SpaceX Starlink 340 7518 340 -18.3 -36.6 2.5% 190.3 -23.0 0.2% 12.6 1.51 3.2 4.8 550 1600 550 -23.0 -46.0 4.0% 63.5 -30.1 0.4% 6.2 1.58 4.2 5.9 1150 2800 1150 -32.1 -64.2 7.6% 214.0 -44.9 1.2% 34.8 1.79 5.8 7.5 OneWeb 648 1200 -32.7 -65.4 7.9% 51.3 -45.9 1.3% 8.6 1.81 5.9 7.5 Amazon Kuiper 590 784 590 -23.8 -47.5 4.2% 33.2 -31.3 0.4% 3.4 1.60 4.4 6.0 610 1296 610 -24.1 -48.2 4.4% 56.6 -31.9 0.5% 6.0 1.60 4.5 6.1 630 1156 630 -24.5 -49.0 4.5% 52.0 -32.5 0.5% 5.6 1.61 4.5 6.1 Sat Revolution 1024 350 -18.6 -37.1 2.6% 26.6 -23.4 0.2% 1.8 1.52 3.2 4.9 China CASC 320 1100 -31.5 -63.0 7.4% 23.5 -43.9 1.2% 3.7 1.78 5.7 7.4 China LuckyStar 156 1000 -30.2 -60.4 6.8% 10.6 -41.7 1.0% 1.6 1.74 5.5 7.2 China Commsat 800 600 -23.9 -47.9 4.3% 34.4 -31.6 0.4% 3.6 1.60 4.4 6.0 China Xinwei 32 600 -23.9 -47.9 4.3% 1.4 -31.6 0.4% 0.1 1.60 4.4 6.0 India AstroTech 600 1400 -34.9 -69.8 9.0% 54.0 -49.7 1.6% 9.9 1.88 6.3 7.9 Boing 2956 1030 -30.6 -61.2 7.0% 205.6 -42.4 1.1% 31.2 1.75 5.6 7.2 LeoSat 108 1423 -35.2 -70.3 9.1% 9.9 -50.1 1.7% 1.8 1.89 6.3 7.9 Samsung 4700 2000 -40.4 -80.9 11.9% 561.2 -59.2 2.7% 124.8 2.11 7.0 8.7 Yaliny 135 600 -23.9 -47.9 4.3% 5.8 -31.6 0.4% 0.6 1.60 4.4 6.0 Telesat LEO 117 1000 -30.2 -60.4 6.8% 7.9 -41.7 1.0% 1.2 1.74 5.5 7.2 (Iridium) 66 780 -27.0 -54.0 5.5% 3.6 -36.5 0.7% 0.5 1.66 5.0 6.6 Total 26,750 1,605 258 Table 1. List of the constellations used in this study; it is meant to provide a representative sample. The list includes the number of satellites (column 1) and their altitude in km (column 2). Column 3 is the elevation of the Sun for which a satellite at Zenith is just illuminated, also for which half the satellites in range are illuminated; by construction, it is also the angle Zenith-Centre of Earth-Satellite at which the satellite is on the horizon. Considering the cases of satellites above the horizon and above a zenithal distance of 60◦ , the table also lists the minimum solar elevation required to illuminate all these satellites (Columns 4 and 7); the fraction of the constellation that is in range (Columns 5 and 8) and the corresponding number of satellites in range (columns 6 9). For reference, the orbital period is listed (column 10), as well as the magnitude of a satellite at zenith (column 11) and at z = 60◦ (column 12), for an intermediate solar phase angle of 90◦ , considering our simple photometric model. Using the right triangle SCH, we have The value of N is reported for each constellation in Table 1, for RSat − hz hz zmax = 90◦ (objects above the horizon) and zmax = 60◦ . cos γ = =1− , (6) Using these formulae, Figure 3 shows the number of constel- RSat RSat lations in range above a range of elevations, and the correspond- where hz is the length of the segment HZ, or the height of the ing density of satellites per square degree for the considered con- zenithal cap defined by the position of the satellite at its zenithal stellations. It also displays the fraction of constellations in range distance z. Inserting the expression of γ from Eq. 5 into Eq. 6, as a function of elevation for various satellite altitudes; this plot we obtain the expression for hz : can be used to estimate numbers for arbitrary constellations. It is hz = (1 − cos γ)RSat , interesting to note that while 2–12% of a constellation is above REarth !!! the horizon, this fraction drops to 1–6% at 15◦ elevation, and = RSat 1 − cos z − arcsin sin z . 0.5–3% at 30◦ , for altitudes in the range of 250–2000km. RSat (7) Galadí-Enríquez (2019) performed detailed simulations of Now that we have hz , we can compute the area of the spheri- the number of satellites in range from various observatories us- cal cap above z, ing the actual orbital distribution of the Starlink satellites as Walker constellations, and obtained N ∼ 40 – 80 with z < Avis. = 2π RSat hz , (8) 60◦ , from low- to high-latitude observatories, respectively. The and the area of the whole sphere containing the satellite, present geometric approximation finds 54 satellites in the same Atotal = 4π R2Sat . (9) conditions (independently of latitude). The referee of the present paper also independently obtained similar numbers with a simi- The number of satellites visible in the cap above z is therefore lar method. AVis. N = NConst. , ATotal 3. Illumination of the satellites hz = NConst. , In the visible and NIR, the satellites are visible only because of 2RSat !!! reflected sunlight. Therefore, in order to be observable, a satel- NConst. REarth = 1 − cos z − arcsin sin z . lite must be both in range (above the horizon, or above z) and 2 RSat (10) illuminated. Article number, page 3 of 12
A&A proofs: manuscript no. output Zenith Z S hz Satellite β H h e z δ O Horizon a Elevation of Sun Observatory to the Sun a RSat γ Earth REarth h C Centre of the Earth Fig. 2. Angle and vector definitions. In this figure, the altitude of the satellite is h ∼ 1000 km and the Sun is at the lowest elevation that illuminates the satellite. The fraction of visible satellites that are illuminated by the we have Sun varies with the Sun elevation below the local horizon. When Z X the Sun is above or on the local horizon, all the satellites above Ared = cos(arcsin(x))dx the horizon are illuminated. By construction, the angle γo intro- 0 duced above is also the elevation of the Sun below the horizon, 1h √ iX just illuminating a satellite at zenith (i.e. when half the satellites = x 1 − x2 + arcsin x 2 0 in range are illuminated), and 2γo is the elevation of the Sun 1 √ below the horizon, just illuminating a satellite on the horizon to- = X 1 − X 2 + arcsin X . (13) 2 ward the Sun (i.e. the elevation of the Sun below which no satel- lite is illuminated). These values of the Sun’s altitude are listed Using this in the equation for Ailluminated , and with Atotal = π, we in Table 1. Generalising, let a be the elevation of the sun below have the horizon that just illuminates a satellite at zenithal distance z. √ By construction, 1 X 1 − X 2 + arcsin X f = − . (14) 2 π a = γ + γo , (11) The function in Eq. 14 is displayed at Fig. 5. It is approxi- mated by a linear function as and the expressions for γ(z) and γo are given in Eqs. 5 and 1. fIlluminated = 100% if a ≥ 0 Simplifying the geometry of the terminator to make it a a straight line (acceptable for low earth orbit (LEO) satellites as =1− if 0 ≥ a ≥ a(zmax ) a(zmax ) the spherical cap considered is small compared to the sphere), the fraction is obtained by integrating the fraction of the spheri- = 0 if a < a(zmax ) , (15) cal cap on the Sun-side of the terminator, approximated by where a is the actual elevation of the Sun. Figure 6 shows, as a function of the altitude of the satellites, a = γ0 correspond- Ailluminated f = . (12) ing to 50% of the illuminated satellites, and a corresponding Atotal to the whole constellation in range being in the Earth’s shadow, considering the case of z = 60◦ and z = 90◦ (horizon). These For satellites on higher orbits than those considered here, and fractions are applied to all the constellations for various sun ele- for observatories at high latitudes and observations far from the vations from sunset until beyond the time when all satellites are equinox, the actual shape of the shadow cone of the Earth will in shadow. The results are displayed in Fig. 7 for each constella- not satisfy this approximation, but it is acceptable for LEOs. tion and for the complete collection. From Fig. 4, Ailluminated = π/2 + 2Ared where Ared is the area high- The correspondence between the elevation of the sun and lighted in red. With θ the angle from the centre so that X = sin θ, time is represented in Fig. 7.c. Thanks to the uniform satellite Article number, page 4 of 12
Hainaut O. R.: Satellite Constellations Night A N C Day 1600 All constellations 0.25 All constellations 1.00 Fraction of illuminated satellites 1400 0.90 Total number above elevation Density at elevation [sat/deg] 0.20 0.80 1200 1000 0.70 0.15 0.60 800 0.10 0.50 600 0.40 400 0.05 0.30 200 0.20 0 0.00 0.10 0 15 30 45 60 75 0 15 30 45 60 75 0.00 Elevation [deg] Elevation [deg] a -0.10 -72 -66 -60 -54 -48 -42 -36 -30 -24 -18 -12 -6 0 6 12 Fraction of constellation in range Sun Altitude [deg] 0.12 250km 500km Fig. 5. Fraction of the satellites in range and illuminated by the Sun 0.10 750km as a function of the Sun’s elevation above the horizon. The satellites’ total fraction above elevation 1000km orbits have an altitude of 1000 km in this example. The region shaded 0.08 1500km 2000km in dark red corresponds to day time; in red (C) to civil twilight, when 0.06 observations are not possible or are not affected by bright sources; in orange (N) to nautical twilight, when most observations are not possi- 0.04 ble; in yellow (A) to astronomical twilight, when observations in the IR or short observations in the visible are possible. 0.02 0.00 0 15 30 45 60 75 Elevation [deg] 0 b 10 Fig. 3. (a) Left: Number of satellites above a given elevation; Right: Corresponding density (in satellite per square degree) at a given ele- 20 Sun elevation [deg] vation, for all the constellations considered. The total value is in red. 30 (b): Fraction of a constellation above a given elevation for a series of 40 altitudes. 50 60 70 0, 50% illuminated 80 Last illuminated above z=600 Last illuminated above horizon 90 250 500 750 1000 1250 1500 1750 2000 Illuminated satellites sat. Satellite altitude [km] in Fig. 6. Elevation of the Sun at which half the constellation is in shadow darkness (dashed), and below which the whole constellation is in shadow, consid- -1 0 x 1 ering only those above z = 60◦ (orange) or above the horizon (green). The dotted line marks −18◦ , i.e. astronomical twilight. ter author for low- to mid-latitude observatories are small, con- firming the validity of our simplifying approximations. theta It is worth noting that satellites on very low orbits are illu- minated during only a brief period immediately after sunset and before sunrise. Because of this, the trains of satellites on their very low transfer orbit immediately after launch are visible only briefly during twilight. Fig. 4. Top view of the sky visible from an observatory; the yellow area indicates the fraction of the sky (for a given orbit altitude) where the satellites are illuminated by the Sun; the area in red is used to compute 4. Magnitude and brightness of the satellites the area in yellow. The satellites are complex objects, with a complicated reflection and diffusion behaviour. Flat, polished panels (such as some of distribution, these fractions are valid for any observatory, and for the satellite’s body or the solar panels) act as mirrors, causing any date. The linear approximation for the terminator restricts specular reflections that, when pointing toward Earth, affect a their validity to satellites on orbits below a few thousand kilo- very small area of the planet but can cause an extremely bright metres, which is valid for the constellations considered. Galadí- flash. Other parts of the satellites will diffuse light. Furthermore, Enríquez (2019) computed the fraction of illuminated Starlink the satellite attitude with respect to the Sun and the observer will satellites using actual Walker constellations; our results are in complicate matters. In the spirit of this paper, this is simplified agreement with his. The seasonal effects determined by this lat- using a straightforward model of this complex situation. Article number, page 5 of 12
A&A proofs: manuscript no. output Number of illuminated satellites above horizon 1,000 100 Number of Satellites 10 1 0 -84 -78 -72 -66 -60 -54 -48 -42 -36 -30 -24 -18 -12 -6 0 Sun Elevation [deg] Number of illuminated satellites above zMax 1,000 100 Number of Satellites 10 1 0 -84 -78 -72 -66 -60 -54 -48 -42 -36 -30 -24 -18 -12 -6 0 Sun Elevation [deg] All SpX Starlink SpX Starlink 340 SpX Starlink 550 SpX Starlink 1150 OneWeb Amazon Kuiper 590 Amazon Kuiper 610 Amazon Kuiper 630 Sat Revolution China CASC China Lucky Star China Commsat China Xinwei India AstromeTech Boing LeoSat Samsung Yaliny (Iridium) Total Naked Eye −18◦ (astronomical twilight), even fewer observation types (and shorter ones) are possible for > −12◦ (nautical twilight), and virtually none for > −6◦ (civil twilight). The bottom plot (c) indicates the number of hours before and after sunrise and sunset corresponding to the sun altitude, for three observatories, for both solstices and an equinox. 4.1. Visible magnitude and diffusion sphere is described by the phase function: The satellite is simplistically represented by a sphere, charac- terised by its radius r and its albedo p. We consider only simple geometric diffusion (i.e the diffused light is proportional to the 1 + cos α cross-section of the object), and the solar phase effect for the f (α) = , (16) 2 Article number, page 6 of 12
Hainaut O. R.: Satellite Constellations where α is the solar phase angle. The magnitude of the object is then 0 Alt. 300km Alt. 1200km Eye limit 6 1 Alt. 500km Alt. 1400km Eye limit 5 M = MSun − 2.5 log( f (α)r p) + 5 log(R∆) + xχ , 2 (17) 2 Alt. 1000km Alt. 2000km Satellite magnitude [V] where MSun = −26.75 is the magnitude of the Sun (in the V 3 band, around 550nm); r is the radius of the object expressed in 4 astronomical units (1 au = 1.4959791011 m); R is the heliocen- 5 tric distance of the object, 1 au; ∆ is the distance between the 6 object and the observatory, also expressed in astronomical units. 7 The term xχ represents the absorption by the atmosphere, where 8 x ' 0.12 mag/airmass is the extinction (0.12 is a typical value in 9 the visible filter V; see e.g. Patat et al. (2011)) and χ = 1/ cos(z) 10 is the airmass, which is the quantity of atmosphere crossed by 0 15 30 45 60 75 90 Zenithal distance [deg] the observed light, normalized to zenith. This equation is cus- tomary for the magnitude of asteroids; see for example Gehrels Fig. 8. Visible magnitude of Space-X-like satellites as a function of their & Tedesco (1979). Here, ∆ is obtained from the zenithal distance zenithal distance, for various orbit altitudes. Only objects with a mag- of the satellite, nitude above the dashed line can be seen with the unaided eye in good ! conditions (mag 5) and above the dotted line in exceptional conditions REarth − RSat z = arctan cot(γ/2) + π/2 + γ/2 , (18) (mag 6). REarth + RSat in sin γ ∆ = RSat . (19) sin z The radius r and albedo p of the satellite are difficult to es- timate. Measurements of NOAA satellites (1500kg, 3.7×1.88, mag 4.1 at zenith) indicate that r = 1.5m and p = 0.25 reproduce the brightness of the satellite well. Scaling down to the Starlink satellite (550kg) we use r = 1m and p = 0.25. This results in a range of 4.2 to 5.9 mag for Starlink 550km, which is in agree- ment with a direct photometric measurement of V = 5 for such a satellite (T.Tyson, priv. comm.). With this assumption, only objects at the lowest altitudes are visible to the naked eye. The corresponding magnitudes are displayed in Fig. 8. More recent measurements of the Starlink satellites on their final altitude and attitude indicate they could be as faint as ∼ 8 mag; furthermore, Fig. 9. Number of flares for each constellation, simply scaling them to Starlink is experimenting with a darkened coating that could one-third of the flares caused by the original Iridium satellites (which make the satellites even darker. We keep the above-mentioned had three large antennas) and to the number of satellites. This is the number of observable flares per night, or the number of flares per week estimate as a conservative, brighter limit, also accounting for the brighter than −5 mag for a mid-latitude site. The colour encodes the sun fact that other satellites could be brighter than those of Starlink. elevation below the horizon, from 0◦ (red), −18◦ (pale blue), and into Using these values, the post-launch low-altitude SpaceX the night (darker blue to greys). Starlink satellites would appear between −2 and −1 mag, in good agreement with the numerous spotting of the Starlink trains. These bright magnitudes combined with the spectacular ‘string cause noticeable flares. It is unknown which, if any, of the new of pearls’ appearance of these trains explain the attention they satellites in the LEO constellations will cause flares. As a con- have received. servatively pessimistic approach, we consider that every satellite With these assumptions, and considering that all the satellites will have one Iridium-like reflecting surface that causes flares have the same characteristics, the magnitudes of the satellites are similar in brightness and frequency to those caused by Iridium’s listed in Table 1 for an observation at zenith and at z = 60◦ . The antennas. This is extremely pessimistically conservative. Sim- total number of objects in range, illuminated by the Sun, and ply scaling the flare frequencies to one-third of those caused by visible with the naked eye (mag < 5 and < 6) is displayed as a Iridium (where each satellite had three antennas) and to the num- function of the elevation of the Sun in Fig. 7(a and b). ber of satellites leads to a total of about 660 flares visible above the horizon per night including 100 flares brighter than −5 mag. Above z = 60◦ , these numbers convert to 100 flares per night, in- 4.2. Specular flares cluding 20 brighter than mag −5. Assuming that the flares occur The original Iridium constellation was well known for its spec- at random times while the satellite is illuminated, Fig. 9 displays tacularly bright flares, where each of their three ∼ 1 × 2m an- the contribution of each constellation for different solar eleva- tennas illuminated a ∼10 km diameter on the ground. With the tions. 66 satellites on 800 km altitude orbits, Iridium flares were vis- ible quite often (two to four times per night). Flares of -5 mag 4.3. Thermal infrared emission in brightness occurred three to four times per week; flares of -8 mag may be visible three to five times per month for sta- In the 5–20µm range, the satellites will emit a considerable tionary observers (Wikipedia). Newer Iridium satellites do not amount of thermal IR radiation. Simulating the details of that Article number, page 7 of 12
A&A proofs: manuscript no. output emission would not be simple: the surface materials are cho- a b 90 sen to maintain the temperature of the satellite within the opera- Effective 80 4 Apparent Angular velocity at Zenith [deg/min] tional range, and include thermal radiators designed to eliminate the heat generated by the instrumentation and received from the 70 Sun and from the Earth (the Earth-facing side receives significant 6 60 thermal radiation from Earth due to the large viewing angle). Magnitude 50 8 Simplifying this to the extreme, a satellite is represented by a sphere with a diameter of 1m and an albedo of 0.25. An emis- 40 10 sivity of 0.1 leads to temperatures over 400K, probably not re- 30 alistic for the hardware. Using an emissivity of 0.9 leads to an 12 isothermal temperature of ∼ 300 K; for a satellite at an altitude 20 of 2000 km, this would produce a flux of up to 100 Jy in N-band 10 14 (8–13 µm), and several tens of Janskys in the M- and Q-bands (5 500 1000 1500 2000 500 1000 1500 2000 Altitude [km] Altitude [km] and 18-20 µm, respectively; Th. Mueller, priv. comm.). As the satellite operators are certainly striving to keep the c d inside of the satellite at ∼ 300 K, and as the satellites alternate 12.0 300.0km between solar illumination and Earth shadow on an hourly ba- 80 500.0km sis, we consider that the temperature of the satellite is constant, 1000.0km 12.5 Angular velocity [deg/min] 1200.0km and that the thermal IR flux is constant at 100 Jy in N-band and 1400.0km Effective Magnitude 60 50 Jy in M- and Q-bands. These are conservative estimates –the 2000.0km 13.0 actual flux could be significantly lower. The number of satellites 40 relevant for thermal IR observations is then simply the number 13.5 of satellites in range; whether or not they are illuminated by the Sun is irrelevant. 20 14.0 0 5. Observation contamination 14.5 0 25 50 75 25 50 75 Zenithal distance [deg] Zenithal distance [deg] One way to evaluate the fraction of observations affected by satellites would be to compute the position of all the satellites Fig. 10. (a) Angular velocity of a satellite as a function of its altitude, in the sky above an observatory at a given time, ‘shoot’ a series around Zenith; (b) Again at Zenith, effect of the altitude on the apparent of exposures, and compute how many of these have a satellite magnitude of the satellite, and on the effective magnitude accounting for in the field of view. We instead used a geometrical probabilistic trailing (with a seeing of 1 arcsec); (c) Angular velocity of a satellite as approach: considering the duration of an observation (the indi- a function of its zenithal distance, for various altitudes; (d) Combining vidual exposure time), we estimate the fraction of the sky that the effect of distance and trailing on the effective magnitude, showing is covered by satellite trails during that exposure time. The field no dependency with zenithal distance. of view of the observation is accounted for by setting the width of the satellite trails. In that way, we immediately have an es- timate of the probability of having an exposure affected by a 22 satellite: this is estimated as the fraction of the sky covered by 1.0s 3.0s satellite trails. We consider various types of scientific exposure 20 10.0s Effective apparent magnitude times over a representative set of ESO instruments. 30.0s 100.0s 18 300.0s – Standard imaging in the visible (e.g. with FORS2 or 1000.0s EFOSC2), or the NIR (e.g. with HAWKI). Individual expo- 16 sure times range from a few seconds to a couple of minutes for broad-band filters, and to several minutes in narrow-band 14 filters. For the simulation, we use an exposure time of 100s, and a field of view of 60 in diameter. 12 – Wide-field imaging in the visible (e.g. with OmegaCam) or 250 500 750 1000 1250 1500 1750 2000 in the NIR (e.g. with VIRCAM). Exposure times are similar Satellite Altitude [km] as in the previous case. We use 100s, and a field of view of 1◦ . Fig. 11. Effective apparent magnitude of the satellite as a function of its altitude, for various exposure times. A field star with that magnitude – Long-slit spectroscopy, in the visible (e.g. with FORS2) or will have the same peak brightness as the satellite in that exposure. the NIR. Typical exposures range from a few minutes to one hour. We use 1000s, and a slit length of 60 . – Short-slit spectroscopy, in the visible (e.g. with UVES – Multi-fibre spectroscopy, for example with FLAMES or or XSHOOTER) or the NIR (e.g. with CRIRES+ or 4MOST. Typical exposures range from a few minutes to one XSHOOTER). Typical exposures range from a few minutes hour. The fibres are positioned over a broad but very sparsely to one hour. We use 1000s, and a slit length of 12”. populated field of view. 4MOST has 2400 fibres on a 4.6 – Fibre-fed spectroscopy in the visible (e.g. with HARPS or sq.deg field of view with a 2.6◦ diameter. In the worst case, ESPRESSO) or the NIR (e.g. with NIRPS). Typical expo- up to 30 fibres could be affected by a satellite trail. The ef- sures range from a few minutes to one hour. We use 1000s, fect on this instrument is obtained by multiplying the effect and a fibre diameter of 2”. on one fibre by 30. Article number, page 8 of 12
Hainaut O. R.: Satellite Constellations For spectrographs, we consider that the slit is always perpen- a pessimistic limit: a faint satellite would potentially affect only dicular to the motion of the satellite, conservatively maximising a quadrant of the camera. The estimates presented below scale the cross-section. with the exposure time and the field of view, and therefore the To estimate the length of the trail left by a satellite during effects can be adjusted to other instruments and exposure times. an exposure, the observed angular velocity is obtained by com- The area A of the sky covered by satellite trails is obtained puting numerically the derivative of the zenithal distance (from from A = tvwN, where t is the exposure duration in seconds, v Eq. 18) accounting for the orbital velocity of the satellite using the angular velocity of the considered satellites in deg/second, Kepler’s law. w the width of the field of view for the considered observation The apparent angular velocity is a function of the altitude type in degrees, and N the number of considered satellites in (an object being further away appears to move more slowly be- range and illuminated. The contributions of the various types of cause of the slower intrinsic motion and the larger distance; satellites are summed, resulting in the total contaminated area. see Fig. 10(a)), and of the zenithal distance (the lower the ob- As N is a function of the elevation of the Sun below the hori- ject, the slower its apparent motion because of foreshortening; zon, the computation is repeated for various bins of solar eleva- Fig. 10(c)). The effective magnitude of a satellite will depend tion. If the length of a trail tv is too long to fit in the observable on the distance between the satellite and the observer, which is sky, this simply means that the first satellite disappeared over a function of the zenithal distance of the satellite, and on the the horizon, and was replaced by a new one entering the observ- trailing of the satellite during the exposure, which is a func- able sky. The total area of the observable sky above z = 60◦ tion of its angular velocity. For a typical seeing of 1 arcsec, the is Asky = (1 − cos(60◦ )/2 × 41 252.96 sq.deg = 13 323 sq.deg. length of the trail in arcsec will give the attenuation factor in Overlapping satellite trails are counted separately, resulting in an magnitude, 2.5 log(v) (v in arcsec/sec). The geometric effect and overestimation of the contaminated area (ultimately, this could trailing attenuation effect are illustrated in Fig. 10(b). The addi- result in an estimated contaminated fraction > 100%). tional effects of the zenithal distance on the geometric distance In the case of flares, it is assumed that the duration of the and on the apparent velocity compensate each-other, as seen in flare is t = 10 s. The number of (10s) flares at a given time is Fig. 10(d). As a reference, Fig. 11 displays the apparent effective computed scaling the frequencies (in number of flare by night) magnitude of the satellite as a function of its altitude, for various to the duration of a flare, accounting for one night, which is equal exposure durations. to 10h or 3600 times the duration of a flare. The width of the contamination trail depends on the magni- In the case of thermal IR emission, the effect of the satel- tude of the satellite and the observing technique: lites does not depend on them being illuminated or not, so the contaminated fraction of the sky does not change with solar ele- – Bright flare (mag < 0): The satellite trail heavily saturates vation. the detector. We consider that the whole field of view of the The contaminated fraction directly gives the probability that instrument is contaminated by the trail, either directly, by a given exposure will be lost due to a satellite; these are listed spurious reflections and diffusion of the light from the trail in Table 2. We note that these fractions scale linearly with the in the instrument, or possibly by contamination caused by exposure times and the field of view, meaning that the effect on cross-talk or interference in the saturated electronics of the other specific exposures can be inferred from this table. instrument. – Medium-brightness satellites (0 5) leaving a non-saturated trail on the detector: the track width will extend over a few times Even considering very pessimistic estimates (i.e. each satellite the seeing, conservatively set to 5” for imagers and long-slit has one Irridium-like reflecting surface) and considering the spectroscopy, and to the full slit in case of a short slit. complete collection of constellations with their conservatively pessimistic uniform distribution, only long exposures (1800 s) The boundary values between the brightness categories are with wide-field survey cameras (1 sq.deg) would be marginally a function of the diameter of the telescope and of the sensitiv- affected (at the 10−4 level). One must note that this type of ex- ity, dispersion, and transmission of the optical elements: a low- posure is extremely rare (images have rarely an exposure time efficiency spectrograph on a small telescope will indeed be less longer than 10 min). All the other categories of exposures are affected than an imaging camera with broad-band filters on a gi- affected much below the 10−4 level. Specular flares are therefore ant telescope. The values chosen are representative of a large not considered an issue for telescopic astronomical observations. telescope like ESO’s 8m Very Large Telescope (VLT), but are also valid for smaller telescopes like the ESO 3.6m New Tech- nology Telescope and the upcoming 39m Extremely Large Tele- 5.2. Contamination of observations in the visible and scope (ELT). near-infrared by satellite trails Because of the extreme case of the Rubin Observatory (for- The effect of satellite trails is different for the various types of merly known as LSST, with a large diameter of 8m, high exposures considered: sensitivity of the detectors, and gigantic field of view of 10 sq.degrees), we consider it separately. Based on reports by – Short exposures (1s) are essentially not affected by the Tyson (priv.comm.), the effect of a bright satellite trail contam- satellite trails. inates the full field of view, and that of a fainter satellite con- – Medium-duration exposures (100s) are affected at a very taminates a full quadrant of the instrument (either directly, or low level (below 0.1%) during the night, and at a low level through electronic cross-talk in the camera electronics causing (0.5%) during nautical twilight. unremovable signal). LSST observes typically with an exposure – Long spectroscopic exposures (1000s) are affected at less time of 30s. For the simulations presented below, the field diam- than the % level during the first and last couple of hours of eter is set to 3.5◦ for all satellites brightnesses, which constitutes the night, and at the 1% level during astronomical twilight. Article number, page 9 of 12
A&A proofs: manuscript no. output This can –in most cases– be mitigated by not scheduling long the satellite, arcsin( diameter/distance), is of the order of 0.2–1”. exposures during the astronomical twilights (which are usu- The apparent angular velocity of the satellites (see Fig. 10) is of ally not suitable for these observations anyway) and the first the order of 15 to 80 deg/min. and last hour of the night. The occultation duration for a point-source is therefore in the – Wide-field imaging (OMEGACAM) and multi-fibre spec- 2 × 10−4 to 1 × 10−3 s range. Using the same mechanism as in trographs (4MOST) are affected at the 5–7% level at the Sect. 5, the field of view of an occultation is 1 sq.arcsec (i.e. a beginning and end of night. conservative value for the angular size of the satellite), and the – LSST ultra-wide exposures on a large telescope: up to 30% width of the trail is set to 1”. The exposure times considered are of the exposures would be lost during the first and last hours 10, 1, and 0.1 s. Accounting for low- and high-orbit satellites, of the night, and almost 50% of the twilight exposures would the probability of one exposure being affected by an eclipse is be contaminated. The combination of wide field of view and ∼ 10−4 , 0−5 , and 10−6 (respectively). The effect of the eclipse the huge collecting area of a large mirror makes this type of ranges from 2×10−5 mag (10s exp, low satellite) to 1×10−2 mag observation very sensitive to satellites. This is likely to cause (0.1 s, high satellite). significant disruption in the scheduling and efficiency of the Overall, the effect therefore ranges from negligible to small surveys. (10 mmag is about the limit of what can be measured from the – Caveat: Except in the case of the LSST, these estimates con- ground). The probability of these occultations occurring is small: sider that a long-slit spectroscopic frame or an image affected at worst, ∼ 10−4 of 10 s exposures affected, or about one 10 s by a faint satellite (mag fainter than 5, effective magnitude exposure every three nights of observation. below 16) is not completely ruined by the trail, and that the remaining part of the frame can be used for science, for in- stance by combining it with other frames. This will not be 6. Mitigation measures true for all science cases: for some programmes, any trail in Two main types of mitigation can be considered. The first is the field of view could ruin the whole frame, no matter how scheduling of the observations: At the global level, observing faint it is. For these science cases, the fraction of affected toward the direction opposite to the Sun (toward the east in the exposures could be in the 10–20% level around twilight de- evening and toward the west in the morning) will ensure that pending on the exposure time and field of view, and mitiga- the satellites are in the shadow of the Earth, therefore avoiding tion measures would be needed. contamination of the exposures. While this is simple to imple- ment and will work even for a wide field of view, this mitigation method is not suitable for all programmes. At a much more de- 5.3. Contamination of thermal infrared observations tailed level, it is possible to forecast the position of the satellites The signal in ground-based thermal IR observations is domi- from their orbital elements, and to observe a field at a time when nated by the thermal emission of the sky and of the telescope, it will not be crossed by a satellite. The implementation of this the astronomical component being a small addition to that bright mitigation is much more complex, and is not suitable for all pro- background. This requires extremely short individual exposures grammes (e.g. it may turn out to be impossible to schedule a long (a fraction of a second, e.g. 0.02 s for VISIR on the ESO VLT), exposure with a wide field of view). and an observation method using chopping (typically by moving The second type of mitigation involves interruption of the the secondary mirror of the telescope) at a few Hertz, and nod- observations: For programmes that require observations in the ding the whole telescope every few seconds. During one of these region of the sky where the satellites are illuminated, it is possi- 0.02 s individual exposures, a static 100 Jy source would be de- ble to compute the exact time when a satellite will cross the field tected with a signal-to-noise ratio (S/N) of ∼ 100 with VISIR. of view, and close the shutter during that time. The implementa- However, the image of the satellite on the detector will be trailed tion of that mitigation would be complex, and is not suitable for by an amount that depends on its altitude, and the flux will be all programmes (e.g. a large field of view could require so many scaled with the inverse square of the altitude. Overall, a satellite interruptions that the exposure would not be practical). In both would in all cases leave a highly visible trail, with an S/N of 9, cases, the availability of high-precision, up-to-date orbital ele- 12, and 50 for a satellite at 2000, 1200, and 300 km, respectively. ments for all the satellites would be crucial so that the accurate However, these trails are sufficiently faint compared to the bright position and timing of the satellites can be computed. background that they would not have additional side effects. Using again VISIR as an example, with a field of view of 38 × 38”, the very short individual exposure time results in ex- 7. Summary tremely low probabilities that an individual exposure will be af- This study presents a very simple evaluation of the effect of fected by a satellite trail (about 10−6 during civil twilight). How- mega-constellations of low-altitude satellites on telescopic as- ever, in most observing modes the individual exposures are not tronomical observations in the visible and IR wavelength do- saved separately but are combined, averaging all the data ac- mains. The main simplifications are (i) a uniform distribution quired on one nodding position, resulting in ∼ 10s. The prob- of the satellites over the globe, (ii) a simple –but empirically ability that at least one of the individual exposures composing calibrated– model for the brightness of the objects, and (iii) a ge- that average is contaminated is of the order of 0.1% during civil ometric probabilistic approach of the contamination. Because of twilight. Overall, thermal IR observations are therefore not sig- the very drastic simplifications of the problem, its results have to nificantly affected by the emission of the satellites. be considered as order-of-magnitude estimates, and will need to be refined using detailed simulations including the actual satel- 5.4. Occultations lite orbits, a refined photometric model of the satellites (ideally tuned for the various satellite models across constellations), a When a non-illuminated satellite passes in front of an astronom- less crude description of the effect of a satellite trail on the data, ical source, it will briefly occult the light. The angular size of and so on. Nevertheless, as most approximations are conserva- Article number, page 10 of 12
Hainaut O. R.: Satellite Constellations Table 2. Probability that an exposure is ruined by a satellite trail, expressed as the fraction of the observable sky (down to z = 60◦ ) contaminated by at least one trail during the duration of the exposure, for the considered observing technique. These are listed as a function of the elevation of the sun (in degrees below the horizon). Low- and high-altitude, and bright and faint satellites are evaluated separately and combined in these totals. Various observing techniques are considered, each having a different field of view and typical exposure time. In the case of LSST, because of the heavy saturation of the satellite trails, it is assumed that the whole field of view is entirely ruined by a satellite. tive, and as the number of satellites considered is very large, the – Medium-duration exposures (100 s) with traditional fields of presented results are likely to err on the pessimistic side. view are affected at a very low level during the astronomical This study considers only the visible and IR regimes. A sep- night. Up to 0.5% of imaging observations would be ruined arate paper will deal with the millimetre and submillimetre do- during the twilights. mains. The radio domain is also to be considered separately. – Long exposures (1000s) with long-slit spectrographs: 0.3 to Keeping in mind the limitations of this study, one can already 0.4% of the exposures would be ruined during the beginning draw the following conclusions for when the 26 000 satellites and end of night, and up to 3% of the exposures taken during from 18 representative constellations are launched and are in op- twilight would be rendered useless. Short-slit and fibre-fed eration: instruments are less affected. – Wide-field imaging and spectroscopic surveys: 1–5% of the – About 1600 satellites will be in range (over the horizon) of exposures would be ruined during the beginning and end of an observatory at mid-latitude. Among those about 250 will night, and at a higher level during twilight. be above an elevation of 30◦ above the horizon (i.e. in the – Very wide-field imaging observations on large telescopes part of the sky where observations take place). At the end of (such as those of the Vera C. Rubin Observatory), for which the evening, that is, in astronomical twilight, or at the begin- saturation and ghosting caused by a satellite will ruin the full ning of the morning, astronomical twilight (i.e. when the sky exposure, would be severely affected: about 30% of the expo- is dark for deep astronomical observations), the number of sures could be ruined at the beginning and end of the night. illuminated satellites will be around 1100 above the horizon, The situation is even worse during twilight (about 50% of and 150 above 30◦ of elevation. Of these, about 260 satel- ruined images during astronomical twilight). Rubin observa- lites will be bright enough to be visible with the naked eye tory published a dedicated report based on an independent in exceptional conditions (mag 6 or brighter); about 110 in study (with different assumptions) indicating “a 40% impact good conditions (mag 5 or brighter). Most of them will be on twilight observing time” (Rubin Observatory Project Sci- near the horizon, with up to about 10 above 30◦ of eleva- ence Team 2020). Only nights in the middle of winter would tion –contrary to claims published online that “satellites will be completely unaffected. outnumber the visible stars”. These numbers plummet as the Sun drops further below the horizon. This paper provides a first quantitative estimate of low-orbit – The trains of satellites, forming a bright ‘string of pearls’, satellite constellations on visible, NIR and thermal-IR astronom- brightly visible right after launch, are not an issue for tele- ical observations, showing the key areas where follow-up as- scopic observations: while they are spectacular, they are very sessments are needed and where collaborative efforts between short-lived and visible only briefly after sunset or before sun- the astronomy community, industry, and governments should fo- rise. cus. The results suggest that large telescopes like ESO’s VLT – Specular flares, while potentially spectacular (Iridium’s ones and upcoming ELT will only be moderately affected, although could reach mag -8), are rare and short enough so that their some science cases may require the implementation of mitiga- effect on telescopic observations will be negligible even ac- tion measures, such as scheduling of the observations or inter- counting very pessimistically for one reflecting surface per ruption of the exposures to allow a satellite cross the field of satellite. The occultation of an astronomical source by a view. These mitigation measures have limitations, in particular passing satellite has a very low probability of occurrence, for large fields of view. Wide-field surveys, in particular on large and the effect is below the precision of the measurement. telescopes like the Vera Rubin Observatory, will be severely af- – Short telescopic observations (with an exposure time of fected. Given the noted effect on wide-field surveys presented in ∼ 1s) with any technique will essentially be unaffected by this paper, further studies should examine the scientific implica- the satellite trails. Similarly, observations in the thermal IR tions on time-domain astronomy in general, asteroid and comet regimes will be unaffected by the thermal emission of the discovery and observation, planetary defence, and other affected satellites. science cases. Article number, page 11 of 12
A&A proofs: manuscript no. output Acknowledgements. We are very grateful to David Galadí-Enrìquez and Patrick Seitzer for their comments, assistance (including corrections!) and discussion, to Thomas Mueller for providing the thermal IR flux estimates, to Gie Han Tan for is assistance for the ALMA estimates and useful discussions. We are also grateful to Jo Andersen and Jason Spyromilio for useful comments on the manuscript. Finally, many thanks to our guardian angel, the anonymous referee who provided many insightful comments, validated many results, and spotted some errors and inconsistencies in the original manuscript. References Galadí-Enríquez, D. 2019, Private Communication Gehrels, T. & Tedesco, E. F. 1979, AJ, 84, 1079 Patat, F., Moehler, S., O’Brien, K., et al. 2011, A&A, 527, A91 Rubin Observatory Project Science Team. 2020, Impact on Optical Astronomy of LEO Satellite Constellations, Document-33805, Tech. rep., Rubin Obser- vatory Legacy Survey of Space and Time Walker, J. G. 1984, Journal of te British Interplanetary Society, 37, 559 Article number, page 12 of 12
You can also read